Introduction
INTRODUCTION
Membranes of a living cell form permeability barriers between the cell and the environment,
and in addition, carry out several critical biochemical functions. To allow each membranebound compartment to carry out its role, there is considerable protein traffic between
compartments. Proteins must be able to pass across these membrane barriers or, in the case of
membrane proteins, insert into the bilayer. Protein translocation is taking place both in
prokaryotes and eukaryotes. However, in eukaryotic cells this process becomes more
complex, because the eukaryotic cell is compartmentalized by the membranes into organelles.
A high proportion of cytosolically synthesized proteins has to cross one or more cellular
membranes to reach their final destination, either outside the cell or within an intracellular
compartment (Agarraberes and Dice, 2001).
Chloroplasts are the photosynthetic organelles residing in plant cells which also in addition
carry out many other critical functions. They are organelles of endosymbiotic origin that are
believed to have evolved from free-living oxygenic photosynthetic eubacteria (primary
endosymbiosis) (Weeden, 1981). During evolution, they transferred most of their genetic
information to the host nucleus. To perform their function, they therefore have to import those
gene products posttranslationally from the cytosol into the chloroplasts. Chloroplasts have
inherited several protein translocation systems from their ancestors that are still similar in
function and mechanism to those of free-living bacterial cells like Escherichia coli. In this
section, the similarities and differences among those protein translocation systems will be
discussed.
1. Structure of higher plant chloroplasts
Chloroplasts are endosymbiotic organelles with a prokaryotic origin that still exhibit some
structural and functional similarities to prokaryotes (Weeden, 1981). Similar to those of the
gram-negative bacteria, two units of membranes, termed the outer envelope membrane and
9
Introduction
inner envelope membrane, respectively, surround them. Chloroplasts are the most complex
organelles, both structurally and functionally. In addition to the envelope membranes
surrounding the chloroplasts, these organelles contain an extra membrane system, the
thylakoid membrane, on which the light phase of photosynthesis takes place. Therefore,
chloroplasts are divided into at least six distinct regions: outer envelope membrane,
intermembrane space, inner envelope membrane, stroma, thylakoid membrane, and thylakoid
lumen. The chloroplast thylakoid membrane is unique among biological membranes in its
structure and composition. The photosynthetic machinery of thylakoid membranes comprises
at least five multisubunit oligomeric complexes, including the photosystems I and II and their
light harvesting antenna (LHC, light harvesting complex), the cytochrome complex and the
ATP synthase (Andersson and Barber, 1994; Herrmann, 1996). Each complex consists of
approximately 15-30 proteins. The thylakoid membrane is utilized for multiple different
applications, including protein transport. In addition, the thylakoid lumen contains many
proteins that are important for processes like water splitting, electron transport etc.
Chloroplasts still contain residual genomes that encode some of their proteins. They are
transcribed translated within the organelles, by using the organellor protein synthesis
machinery. However, the protein synthesis capacity of these organelles is strongly reduced, as
many genes originally encoded within the organelle genomes have been transferred to the
nucleus of the host cells. At present, a typical chloroplast genome encodes only about 100
proteins. Considering that a typical cyanobacteria contains more than 3000 genes (Kaneko et
al., 1996), several thousand genes were either transferred to the nucleus or lost. Recently, the
genome of the model plant Arabidopsis thaliana was sequenced completely and predicted to
contain 25498 genes (The Arabidopsis Genome Initiative, 2000). According to predictions by
cellular localization programs, up to 14% of the gene products, i.e. about 3,500 proteins, are
likely to have a chloroplast localization (The Arabidopsis Genome Initiative, 2000;
Emanuelsson et al., 2000). Those proteins are thus synthesized in the cytosol, using the
protein synthesis machinery of the host cells. To complete the full function of the organelles,
the proteins have to be targeted into the organelles by protein import machineries located at
10
Introduction
both the outer and inner envelope membrane of chloroplasts. For the biogenesis of those
proteins that are located in the thylakoid membrane or in the thylakoid lumen, import across
the envelope membranes is only the first step. They must be subsequently transported into or
across the thylakoid membrane.
2. Protein import into chloroplasts
2.1. Transit peptide for chloroplast targeting
Nuclear encoded chloroplast proteins are synthesized as precursor proteins in the cytosol with
N-terminal targeting signals termed transit peptides. For the thylakoid located proteins, an
additional signal is required for targeting to the thylakoid membrane or to the thylakoidal
lumen. The transit peptides of thylakoid proteins can be divided into two types: one type
carries only the envelope transit signal; the other type carries the envelope transit signal and a
thylakoid transit signal in tandem and is therefore termed bipartite transit peptide. For those
thylakoid proteins that carry the first type of transit peptides, an internal, uncleaved thylakoid
targeting signal located in the mature part of these proteins is required.
The envelope transit signal, which mediates post-translational import of chloroplast proteins
across the envelope membranes, is located at the N-terminus. Although the envelope transit
peptides have no common sequence motifs and are highly variable in length (from 20 to more
than 120 residues), they share several general features (von Heijne et al., 1989; Claros et al.,
1997): (1) an uncharged N-terminal domain of about 10 amino acids that is terminated by
glycine or proline; (2) a central domain lacking acidic residues but enriched in hydroxylated
amino acids, (3) a C-terminal domain that is enriched in arginines. They do not fold into
secondary or tertiary structure in an aqueous environment, but might form amphipathic βstrands or α-helices in a hydrophobic environment (von Heijne and Nishikawa, 1991; May
and Soll, 1999; Wienk et al., 1999). This flexibility may allow for significant conformational
adaptability to the multiple components involved in protein translocation (Bruce 2001).
11
Introduction
Recently, direct interaction of the transit peptides with the protein translocation machinery
has been reported (Subramanian et al., 2001). Most of the envelope transit peptides have a
carboxyproximal region with a loosely conserved sequence Ile/Val-x-Ala/Cys-Ala (x refers to
any amino acid) at the proteolytic processing site of the stromal processing peptidase (SPP)
(indicated by the arrow) (Gavel and von Heijne, 1990). For most of the thylakoid proteins, the
envelope transit peptides are removed by SPP soon after they reach the stroma. SPP is
composed of two antigenically related proteins with molecular masses of 143 and 145 kDa
(Oblong and Lamppa, 1992).
The C-terminal part of bipartite transit peptides provides the signal for transport into or across
the thylakoid membrane and is termed thylakoid transfer domain or signal peptide. The
thylakoid transfer signals are necessary and sufficient for pathway-specific transport of
different precursor proteins. Therefore, the structure of the thylakoid targeting domains of the
transit peptides are necessarily more complex. The features of the thylakoid transfer domains
from each pathway will be discussed in detail later. Signal peptides usually end with an A-XA motif for cleavage by the thylakoid processing peptidase (TPP) (Kirwin et al., 1987).
2.2. Translocation machinery at the chloroplast envelope membrane
Precursor protein translocation across the outer and inner envelope membrane is mediated by
the translocon at the outer membrane of chloroplasts (Toc) and the translocon at the inner
membrane of chloroplasts (Tic). Only few components of the import machineries have
homologues in prokaryotic cells like Synchocystis PCC 6803 (e.g. Reuman et al., 1999),
suggesting that these components have a prokaryotic origin, although the precise function of
the prokaryotic homologues remains unknown. Translocation via Toc and Tic occurs
simultaneously for most proteins, probably at regions where the outer and inner membranes
are in close contact (Chen and Schnell, 1999; May and Soll, 1999). Toc contains several
transmembrane proteins: Toc159, Toc75, Toc34, Toc36, and a recently identified component
Toc64 (Chen and Schnell, 1999; Sohrt and Soll, 2000). Toc34 and Toc159 function as
12
Introduction
receptors for precursor proteins (Kessler et al., 1994), while Toc75 forms the aqueous pore
through which the precursor proteins are translocated (May and Soll, 1999). The diameter of
this pore is only 8-9 Å, suggesting that proteins must be fully unfolded during translocation
into the Toc protein pore (Chen and Schnell, 1999). Tic consists of Tic110, Tic55, Tic22 and
Tic20 (reviewed by Jarvis and Soll, 2002), but the precise organization and function of these
proteins is not clear. Insertion of the precursor proteins into the Tic complex requires ATP
hydrolysis within the stroma (Chen and Schnell, 1999). Several kinds of molecular
chaperones are required for protein transport across the inner membrane, including ClpC and
the chloroplast Hsp70 (cpHsp70) (Marshall et al., 1990; Shanklin et al., 1995). Current
models depict ClpC as the motor driving precursor import (Keegstra and Cline, 1999).
Another chaperone, Cpn60, binds to the precursor proteins and assists in protein refolding in
the stroma (Tsugeki and Nishimura, 1993).
3. Protein export systems – from bacteria to thylakoids
Protein translocation systems are structurally and mechanistically diverse from one membrane
system to another, but can nevertheless be divided into roughly two major types: the export
system and the import system (Schatz and Dobberstein, 1996). Import system transports
proteins into a compartment that is functionally equivalent to, or evolutionarily derived from,
the cytosol. Export system transports proteins from the cytosol to an extracytosolic
compartment. According to this classification, translocation of nuclear-encoded proteins into
mitochondria or chloroplasts is facilitated by import pathways, whereas protein translocation
into the bacterial periplasma and the thylakoidal lumen is facilitated by export pathways. All
export systems are characterized by many common features and are phylogenetically related
to the bacterial secretion systems. Therefore, any knowledge of the bacterial protein transport
systems leads to stimulating investigation on the chloroplast systems, and vice versa.
3.1. Protein translocation systems at the E.coli plasma membrane
13
Introduction
Protein translocation across the bacterial cytoplasmic membrane has been studied extensively
in gram-negative bacteria such as Escherichia coli. Gram-negative bacteria are surrounded by
two membranes, the inner membrane and outer membrane, and therefore possess a functional
periplasm residing between these two membranes. A wide range of proteins with a function in
the periplasmic space or outer membrane has to be transported to their final location. These
proteins are synthesized in the cytoplasm as precursors with a cleavable amino-terminal signal
peptide. Depending on the nature of the precursors, different translocation/secretion pathways
are employed for the transport across the inner membrane (Danese and Silhavy, 1998;
Agarraberes and Dice, 2001).
3.1.1. The Sec-dependent pathway
Most periplasmic proteins and outer membrane proteins are transported across the plasma
membrane via a general secretion pathway, which is characterized by a peripheral ATPase,
SecA, and is therefore designated as the Sec-dependent pathway. Proteins transported by this
pathway are synthesized in cytosol as precursors with an N-terminal signal peptide of 18-26
amino acid residues (von Heijne, 1986; Randall and Hardy, 1998; Chou, 2001). These signal
peptides have three characteristic regions: a positively charged region with alkaline amino
acids at the N-terminus (n-domain), a highly hydrophobic region in the middle (h-domain),
and a polar region containing the signal peptidase cleavage site at the C-terminus (c-domain)
(Berks, 1996). Signal peptides alter the folding properties of the mature part of the precursors
(Park et al., 1988) and are recognized by SecA and often also by a cytosolic chaperone, SecB
(Hartl et al., 1990).
Sec-dependent translocation across the plasma membrane is accomplished by the Sec
translocon. Sec translocon consists of two cytosolic proteins SecA and SecB, and at least six
integral membrane proteins, notably SecY, Sec E, Sec G, Sec D, Sec F and YajC (reviewed
by Driessen et al., 1998). These integral membrane proteins form two distinct trimeric
complexes: SecYEG and SecDFyajC. The first complex forms the protein-conducting
14
Introduction
channel, through which the precursor proteins are inserted into the plasma membrane. SecB
binds to the newly synthesized precursor proteins in the cytosol, mediating the interaction
between preproteins and the translocon. In E. coli, SecA is the most abundant component of
the translocase (Driessen, 1994) and has a balanced cellular distribution between soluble and
membrane-bound states (Cabelli et al., 1991). SecA exhibits high affinity interaction with
SecB/preprotein complexes upon binding to SecYEG (Hartl et al., 1990; Economou and
Wickner, 1994). Hydrolysis of ATP bound to SecA leads to the insertion of the precursor
protein into the SecYEG channel and the release of SecA from the membrane. Reiteration of
the SecA insertion-deinsertion cycle results in translocation of the entire precursor protein
(Economou 1998). ATP is absolutely required for this process, and a requirement for protonmotive force is also common but not universal (Schiebel et al., 1991; Nishiyama et al., 1999).
Not all of the components mentioned above are essential for Sec-dependent translocation.
SecA, SecY and SecE are the minimum entity to perform Sec-dependent protein translocation
in bacteria (Murphy et al., 1995; Prinz et al., 1996), whereas SecG, SecDFyajC are nonessential proteins, required only for maximal rates of protein translocation by regulating
insertion-deinsertion cycles (Pogliano and Beckwith, 1994; Duong and Wickner, 1997).
3.1.2. The SRP-dependent pathway
In E. coli, targeting of several highly hydrophobic inner membrane proteins, such as FtsQ and
leader peptidase, is carried out co-translationally by the signal recognition particle (SRP)
system (MacFarlane and Müller, 1995; der Gier et al., 1996). This system is homologous to
the mammalian SRP system involved in the co-translational targeting of ribosome-nascent
chain complexes (RNCs) to the endoplasmic reticulum (ER) membrane (reviewed by Keenan
et al., 2001). The mammalian SRP system is composed of a soluble cytoplasmic signal
recognition particle (SRP) and a membrane-embedded Sec61 translocon. The mammalian
SRP contains one molecule of RNA (7SL) and a complex of six polypeptides (Walter and
Blobel, 1980; 1983), including a 54 kDa subunit. The bacterial SRP was determined as a
15
Introduction
cytoplasmic ribonucleoprotein complex that consists of a 4.5S RNA molecule and a 48 kDa
protein (Ffh = fifty-four homologue) with homology to the mammalian SRP54 component
(Bernstein et al., 1989; Poritz et al., 1990). In addition, the SRP receptor-like protein FtsY
was also identified (Bernstein et al., 1989; Romisch et al., 1989). FtsY is localized both to the
plasma membrane and to the cytosol. In vitro reconstitution experiment demonstrated that the
Ffh, FtsY and 4.5SRNA are functionally sufficiently to target proteins to the membraneembedded translocon (Koch et al., 1999). Bacterial SRP binds to the RNCs through the
interaction of the SRP with the newly synthesized polypeptide, and the resulting RNCribosome-SRP complex binds subsequently to FtsY in a GTP-dependent manner. The RNCs
are delivered by the SRP and FtsY to the SecYEG protein channel, through which the
polypeptides are inserted into the plasma membrane. Therefore, the Sec- and the SRPdependent mechanisms converge at the translocon (Valent et al., 1998; Neumann-Haefelin et
al., 2000). Remarkably, SecY and SecE show homology to two components of the Sec61
channel at the mammalian ER membrane, the Sec61α and Sec61γ protein, respectively,
indicating that the bacterial Sec/SRP system is phylogenetically related to the mammalian
SRP system (Gorlich et al., 1992).
Recently, a novel and evolutionarily conserved component that plays an essential role in the
SRP-dependent translocation has been identified from E. coli. This plasma membrane protein,
YidC, is a homologue to the mitochondrial protein, Oxa1 (for a review, see Luirink et al.,
2001). Oxa1 is a nuclear-encoded mitochondrial inner membrane protein that presumably
forms a novel translocase in the mitochondrial inner membrane (Herrmann et al., 1997). Oxa1
was shown to be essential for correct integration of a subset of inner membrane proteins
encoded by both the mitochondrial and nuclear genomes (Hell et al., 1998; Hell et al., 2001).
The bacterial YidC was found to be associated with the SecYEG translocon and to interact
with the transmembrane segments of the membrane protein FtsQ during its insertion into the
membrane (Scotti et al., 2000). However, depletion of YidC does not impair transport of SRPdependent proteins such as leader peptidase and ProW (Samuelson et al., 2000). It was
suggested that YidC functions in recognizing transmembrane regions and facilitating their
16
Introduction
lateral movement into the membrane (Scotti et al., 2000; Berk et al., 2001; Urbanus et al.,
2001).
3.1.3. The bacterial TAT pathway
Up to 95% of the periplasmic proteins are transported by the Sec-dependent pathway in an
unfolded formation. However, a subset of periplasmic proteins binds complex cofactors in the
cytoplasm (e.g. iron-sulfur clusters, nickel and iron cofactors), and are thus obliged to fold
prior to translocation (Bogsch et al., 1998). Translocation of these proteins takes place in a
Sec-independent manner. Interestingly, almost all of the precursors of these proteins bear a
characteristic twin-arginine (RR) motif within their signal sequences (Berks, 1996), which is a
unique feature of the ∆pH-dependent protein translocation pathway at the thylakoid
membrane (Chaddock et al., 1995). Accordingly, this pathway was termed TAT (twinarginine translocation) pathway in both the bacterial and the thylakoid system (Sargent et al.,
1998; Dalbey and Robinson, 1999; Berks, 2000). The major feature of the TAT-dependent
mechanism is that this pathway is capable of transporting tightly folded globular proteins
(Rodrigue et al., 1996; Chanal et al., 1998; Santini et al., 1998; and reviewed by Müller and
Klösgen, 2005). The bacterial TAT system is even able to export complexes of proteins
formed in the cytosol by a “hitch-hacker” mechanism (Wu et al., 2000). At least one
experiment has shown that secretion of protein through the TAT system requires the protonmotive force (∆µH +) (Santini et al., 1998).
TAT pathway signal peptides have a similar tripartite organization as Sec signal peptides: a
positively charged n-domain, a middle h-domain and a c-domain. The twin-arginine motif is
located at the end of the n-domain, within a conserved (S/T)-R-R-x-F-L-K sequence motif
(Berks, 1996). The two arginine residues, especially the second one, are invariant and the
other motif residues occur at a frequency of more than 50%. Several experiments have shown
that TAT signal peptides target Sec pathway proteins or foreign proteins to the TATdependent pathway (Cristobal et al., 1999; Thomas et al., 2001), while Sec signal peptides
17
Introduction
direct TAT pathway substrates to the Sec apparatus (Rodrigue et al., 1999), although the
folded structure of the TAT pathway substrates do not allow the full translocation. Thus, in E.
coli, the signal peptide alone mediates mutually exclusive sorting of precursor proteins
between the TAT and Sec pathway.
Identification of the bacterial TAT system components have resulted from homology searches
to the first component of the plant ∆pH-dependent import pathway called Hcf106 (Settles et
al., 1997; Weiner et al., 1998). Three homologous genes to hcf106 were found in the E. coli
genome, notably tatA, tatB and tatE. The tatA and tatB genes are located in one operon
together with tatC and tatD in tandem, whereas the tatE gene forms an independent cistron
(Bogsch et al., 1998; Sargent et al., 1998). Mutagenesis has shown that the TatA/B/E proteins
are essential TAT pathway components, and that TatA and TatE can replace each other in
function (Bogsch et al., 1998; Sargent et al., 1998). Investigation of the tatABCD operon has
led to the identification of a further critical component of this system, TatC. Disruption of the
tatC gene leads to a total block of Tat-dependent export (Bogsch et al., 1998), indicating a
central role for TatC in this pathway. On the other hand, tatD encodes a DNase of cytosolic
localization, which is apparently not involved in Tat-dependent protein transport (Wexler et
al., 2000). All bacterial Tat components are membrane proteins of the plasma membrane.
TatA, TatB and TatE span the plasma membrane once, with an Nout-Cin topology. TatC sapns
the plasma membrane six times, with both the N- and C-terminus on the cytoplasmic side
(Gouffi et al., 2002). It has been shown that TatB and TatC are associated together in
equivalent amount to form complexes of approximately 600 kDa molecular weight, and
variable amount of TatA could be detected in these complexes (Bolhuis et al., 2001; Sargent
et al., 2001; de Leeuw et al., 2002). However, electron microscopy of several of these
complexes did not yield images sufficiently unique to allow for the assignment of the
obtained structures to individual Tat proteins, suggesting a flexible assembly of the TAT
translocase from a varying number of subunits. Overexpressed TatA and TatB are tightly
associated in vivo, and form complexes with a molar ratio from 15+4 to 19+4 when TatC is
also present (Sargent et al., 2001, de Leeuw et al., 2001). Electron microscopy of such
18
Introduction
complex shows a double-layered ring structure, suggesting that TatA is involved in formation
of the protein conducting channel.
3.1.4. YidC-dependent membrane protein insertion
Also several membrane proteins are inserted into the bacterial plasma membrane by a Secindependent mechanism. These proteins include the procoat protein of phage M13 and phage
Pf3. Both proteins are synthesized in a precursor form, with a cleavable N-terminal signal
peptide. The signal peptide consists of a positively charged N-terminus, followed by a
hydrophobic region. Targeting of the protein to the membrane requires basic residues in both
the N- and C-terminal regions, indicating an electrostatic binding of procoat to the acidic
phospholipid head groups (Gallusser and Kuhn, 1990). The hydrophobic regions in the signal
sequence and the mature protein synergistically contribute to drive proteins insertion into the
membrane (Soekarjo et al., 1996). Translocation of the loop between the two hydrophobic
regions is stimulated by the proton motive force (pmf) across the membrane (Kuhn et al.,
1990). Once inserted, the signal sequence is cleaved by the leader peptidase to generate the
mature protein in the membrane (Shen et al., 1991).
M13 procoat is even able to insert into protein-free liposomes (Geller and Wickner, 1985;
Soekarjo et al., 1996). Experiments have also shown that Pf3 procoat is able to insert into a
trypsin-pretreated membrane (Kiefer and Kuhn, 1999). All of these experiments substantiated
the widely held view that the membrane insertion of this kind of proteins is independent of
any protein factors. However, later experiments have revealed the involvement of YidC in the
membrane insertion process. Membrane insertion of the M13 procoat is strongly inhibited in
YidC-deficient E. coli cells, while the secretion of the periplasmic proteins is not affected
(Samuelson et al., 2000). Direct interaction of YidC with M13 procoat and Pf3 procoat was
furthermore observed during membrane insertion (Samuelson et al., 2001; Chen at al., 2002).
This suggests a dual role of YidC in both Sec-dependent and Sec-independent protein
insertion into the bacterial membrane.
19
Introduction
RIB
RIB
RIB
SecA
SRP
SecB
FtsY
Sec
+ATP SRP
+GTP
+ATP
TAT
cytosol
YidC
RIB
TatB
G
YidC
SecYE
SecDFYajC
TatA/E
FtsY
SecYEG
lep
TatC
YidC
SRP
lep
inner membrane
periplasma
lep
outer membrane
Fig. 1. Protein transport pathways in E. coli. The majority of the periplasmic proteins is
synthesized in precursor form with an N-terminal signal peptide and transported across the plasma
membrane by either of two pathways, the Sec-dependent or the Tat-dependent pathway. A subset of
inner membrane proteins is co-translationally inserted into the membrane by the SRP-dependent
pathway through the SecYEG channel. Other inner membrane proteins are inserted by assistance of
YidC. Upon translocation, the signal peptides are removed by leader peptidase (Lep), which has its
activity exposed to the periplasmic side of the inner membrane. Arrows indicate the protein
translocation pathways from the cytosol to the inner membrane and the periplasm. Specific
requirements of energy and characteristic factors of each pathway are indicated. RIB = ribosome.
3.2. Protein translocation systems at the thylakoid membrane
Nearly all proteins of the thylakoid are encoded in the nucleus and synthesized in the cytosol
as precursor proteins. Most of them undergo a two-stage import process: first, chloroplast
envelope translocation and stroma targeting result in a stromal intermediate form generated
after removal of the import signal by SPP; second, thylakoid transport and lumen targeting
lead to the generation of a mature form, generated by cleavage of the intermediate by TPP in
the lumen. Once in the stroma, the precursor proteins enter precursor-specific transport
pathways. To date, four distinct pathways for protein transport across the thylakoid membrane
have been described. In contrast to the protein import systems at the envelope membrane,
20
Introduction
these pathways are directly related to the protein transport pathways operating at the bacterial
plasma membrane. Accordingly, they are defined as Sec-dependent pathway, SRP-dependent
pathway, ∆pH-dependent pathway, and spontaneous insertion pathway, respectively.
3.2.1. The Sec-dependent pathway of chloroplasts
A subset of thylakoid lumen proteins, including the 33-kDa photosystem II protein (33K),
plastocyanin (PC), and PSI-F, are transported by a mechanism that relies on ATP and soluble
stromal factors (Hulford et al., 1994). Thylakoid transport of 33K, PC and PSI-F is sensitive
to sodium azide (Knott and Robinson, 1994; Karnauchov et al., 1994; Henry et al., 1994),
which is known as an inhibitor of the bacterial SecA protein (Oliver et al., 1990), which
suggests already that a conserved protein transport mechanism is functional within
chloroplasts.
i. Components of the thylakoidal Sec-dependent machinery
Preliminary evidence for a Sec-dependent pathway existing in thylakoids includes the
presence of SecA- and SecY-homologous genes in the chloroplast genomes of several algae
(Scaramuzzi et al., 1992; Douglas, 1992). Later on, SecA-homologous genes were identified
from different plants by a homology-based approach (Berghöfer et al., 1995; Nohara et al.,
1995), and were designated cpSecA. CpSecA was also identified independently by a genetic
approach, in which the maize mutant tha1 was found to result from disruption of the maize
Sec-dependent protein translocation (Voelker and Barkan, 1995). CpSecA is localized in the
chloroplast stroma, and pre-incubation of stroma extract with anti-cpSecA antibodies blocked
transport of 33K and PC into thylakoids (Nakai et al., 1994).
A chloroplast SecY homologue was identified in the Arabidopsis EST collection (Laidler et
al., 1995) and led to the isolation also of the homologous cDNA from spinach (Berghöfer,
1998). A gene homologous to bacterial secE gene is also present in the Arabidopsis genome
sequence (Bevan et al., 1998). CpSecE forms a 180 kDa complex together with cpSecY
21
Introduction
(Schuenemann et al., 1999). Homologues to the other components of the bacterial Sec
pathway, notably SecB, SecG, SecD, SecF or YajC, are apparently lacking from the
Arabidopsis proteome. Since in the bacterial Sec-dependent system, the SecYE complex and
SecA form a minimal translocase to perform protein secretion, the plant thylakoid Secdependent system appears to operate with the minimal number of required components. It is
possible though that additional factors, like for example stromal chaperones such as ClpC,
cHsp70 or Cpn60, might be involved in the Sec-dependent translocation in the chloroplast
system.
ii. Mechanism of thylakoidal Sec-dependent protein transport
The homology observed between the bacterial and chloroplast Sec-dependent machineries
suggests that they both operate in a similar manner. Indeed, although the mechanism of the
chloroplast Sec-dependent translocation has been studied only in a limited number of
experiments, it seems to be largely similar to its bacterial counterpart. Thylakoid precursor
proteins bind to the membrane in a cpSecA-dependent manner (Keegstra and Cline, 1999) and
form a complex within the membrane that also contains cpSecY (Mori and Cline, 2001).
CpSecA partially inserts into the lipid bilayer, carrying a fragment of the precursor protein
(Berghöfer, 1998). ATP is absolutely required for thylakoid Sec-dependent translocation,
because depletion of ATP by apyrase treatment completely prevents Sec-dependent
translocation (Hulford et al., 1994). Inhibition of ATP hydrolysis by sodium azide or by the
ATP-analog AMP-PNP results in a permanent insertion of the SecA into the thylakoid
membrane (Berghöfer, 1998), suggesting that ATP hydrolysis is required for the release of
cpSecA upon translocation of the precursor protein. Tightly folded proteins are not
transported by the thylakoid Sec-dependent pathway (Hynds et al., 1998). A trans-membrane
potential is not essential for Sec-dependent transport, although translocation of some
precursor proteins is stimulated by ∆pH across the thylakoid membrane (Yuan and Cline,
1994; Mant et al., 1995).
22
Introduction
3.2.2. The SRP-dependent pathway of chloroplasts
Thylakoid membrane is one of the most condensed membrane containing a large number of
membrane proteins. In the light of the completion of the genome sequencing of Arabidopsis,
nearly 350 proteins are predicated to locate within the thylakoid membrane (Peltier et al.,
2002), most of which are integrated into the thylakoid membrane by the chloroplast SRP
pathway, a counterpart of the bacterial or mammalian SRP pathway. The majority of these
proteins is encoded by the nucleus and is imported into the stroma after being synthesized in
the cytosol. Only a few of them are encoded by the plastid genome and synthesized in the
stroma of chloroplasts. Obviously, co-translational insertion is possible in chloroplast only for
the plastid-encoded thylakoid proteins. The nuclear-encoded proteins that are synthesized in
the cytosol need to be inserted post-translationally into the thylakoid membrane. Chloroplast
SRP (cpSRP) is unique in that it is capable to insert membrane proteins both cotranslationally and post-translationally. So far, understanding of the mechanism of the posttranslational SRP-dependent protein insertion is largely based on studies of the major lightharvesting chlorophyll a/b binding protein (LHCP), which is encoded in the nucleus. Yet, the
mechanism of the co-translational SRP-dependent protein insertion is by far less well studied,
mainly due to technical constrains.
i. Components of the chloroplast SRP-dependent pathway
Identification of a chloroplast SRP54 homologue (cpSRP54) (Franklin and Hoffman, 1993)
and the discovery of the interaction between cpSRP54 and LHCP in the stroma (Hoffman and
Franklin 1994) led to the conclusion that SRP is operating in thylakoid membrane protein
insertion. Further characterization of cpSRP failed to identify an RNA component, and
instead led to the identification of a novel 43 kDa protein subunit (cpSRP43) (Schuenemann
et al., 1998). Unlike cpSRP54, the evolutionary origin of cpSRP43 remains uncertain, since a
prokaryotic homologue was not identified so far. It might have evolved to cope with the
obligatorily post-translational mode of insertion for nuclear-encoded thylakoid membrane
23
Introduction
protein. Reconstitution of a functional transit complex consisting of cpSRP54, cpSRP43 and
LHCP demonstrated that a RNA component, as well as any further stroma factor, is not
required for insertion of LHCP (Schuenemann et al., 1998). A chloroplast homologue of the
SRP receptor, cpFtsY, was identified in Arabidopsis proteome (Kogata et al., 1999). AnticpFtsY antibodies specifically inhibit the integration of LHCP into isolated thylakoids,
indicating that cpFtsY plays an essential role in the insertion process. However, unlike the
bacterial SRP system, the involvement of the SecYE translocon in the insertion of LHCP has
not been proven. Antibodies to SecY, which block the translocation of lumenal proteins via
the Sec translocon, have no effect on the insertion of LHCP (Mori et al., 1999). Instead, Alb3,
the chloroplast homologue to mitochondrial Oxa1 and bacterial YidC proteins, was shown to
be essential for thylakoid membrane insertion of LHCP (see below). Alb3 is located in the
thylakoid membrane. Antibodies raised against Alb3 specifically inhibit the insertion of
LHCP, but have no effect on Sec- and ∆pH/TAT-dependent protein translocation (Moore et
al., 2000).
ii. Mechanism of post-translational insertion of thylakoid membrane proteins by the SRPpathway
Studies on post-translational insertion of thylakoid membrane protein were almost exclusively
performed with a single model protein, notably LHCP. Like most other nuclear-encoded
thylakoid proteins, LHCP carries a cleavable N-terminal import signal, but no cleavable
thylakoid-targeting signal. Instead, the thylakoid targeting information is located within the
mature part of LHCP (Lamppa, 1988; Viitanen et al., 1988). Once in the stroma, LHCP binds
rapidly to cpSRP54 and cpSRP43 to form a transit complex (Payan and Cline, 1991;
Schuenemann et al., 1998). Formation of the SRP/LHCP complex prevents hydrophobic
LHCP from aggregation and misfolding in stroma (Payan and Cline 1991). CpFtsY binds to
the SRP/LHCP transit complex in a strictly GTP-dependent manner to form a super-complex
(Tu et al., 1999). Like their cytoplasmic homologues, both cpSRP54 and cpFtsY are GTPases
(Hoffman and Franklin, 1994).
24
Introduction
It is still unknown how the transit complex is targeted to the thylakoid membrane. It was
suggested that cpFtsY pilots the transit complex to the membrane, like its homologue does in
the bacterial cytosol (Tu et al., 1999). Once at the membrane, the transit complex is
disassociated upon GTP hydrolysis, delivering LHCP to the membrane component (Groves et
al., 2001). As mentioned above, Alb3 is essential for insertion of LHCP. It was suggested that
Alb3 forms a separate translocon, which is independent of the SecYE complex (Moore et al.,
2000; Eichacker and Henry, 2001), indicating that post-translational integration of LHCP
takes place by a mechanism that is quite different from that of co-translational protein export.
iii. Mechanism of co-translational insertion of thylakoid membrane protein
In addition to its role in post-translational protein export, cpSRP appears to have retained cotranslational targeting activity. A subset of thylakoid proteins, including PsaA, PsaB, PsbA
and PsbD, are encoded by plastid genes. In vitro transcription/translation of psbA gene by
extracts from chloroplast stroma has shown that a PsbA-RNC is found, which interacts with
cpSRP54 (Nilsson et al., 1999). Analysis of Arabidopsis mutants has shown that lack of
cpSRP54 has an effect on the membrane insertion of reaction center proteins such as PsaA
and of LHCP, whereas the lack of cpSRP43 affects only the membrane insertion of LHCP but
not that of the reaction center proteins (Jonas-Straube et al., 2001). This suggests that
cpSRP43 is not involved in the co-translational transport of thylakoid membrane proteins.
Surprisingly, co-translational transport of thylakoid membrane proteins was also not affected
in an Arabidopsis mutant lacking cpFtsY (Amin et al., 1999), suggesting that cpFtsY is not
essential for co-translational insertion. In analogy to the bacterial SRP system, cotranslational integration of thylakoid membrane proteins is apparently dependent upon the
SecYE translocon. In a maize cpSecY null mutant, translocation of PsbA was also severely
affected (Voelker et al., 1997; Roy and Barkan, 1998), indicating that cpSecY is required, for
example, for ribosome binding during co-translational transport of plastid-encoded proteins.
The role of Alb3 in co-translational transport, one the other hand, is not yet established. The
thylakoidal ∆pH stimulates the co-translational SRP-dependent export (Zhang et al., 2000;
25
Introduction
Muhlbauer and Eichacker, 1999), as well as the post-translational SRP-dependent membrane
insertion of LHCP (Cline et al., 1992).
3.2.3. The ∆pH/TAT-dependent pathway of chloroplasts
In vitro studies of protein import into isolated thylakoids revealed that a subset of precursor
proteins is transported by a mechanism that is different from the Sec- or SRP-dependent
pathways. These proteins include the 16- and 23-kDa subunits of photosystem II (16K and
23K), photosystem II subunit T (PSII-T), and photosystem I subunit N (PSI-N) (Mould and
Robinson 1991; Mould et al., 1991; Cline et al., 1992; Klösgen et al., 1992; Henry et al.,
1994). Import of those proteins requires neither nucleoside triphosphates nor soluble stromal
factors but is instead totally dependent on the ∆pH across the thylakoid membrane. Therefore,
this pathway was designated the ∆pH-dependent pathway.
i. Pathway selection of precursors: distinctive signal peptides
Competition studies with chemical amounts of precursor proteins showed that the Sec- and
∆pH-dependent pathways are precursor-specific and that they operate independently from
each other in the translocation of thylakoid proteins (Cline et al., 1993). No genuine precursor
targeted by both pathways has so far been found, although recombinant precursors and the
cyanobacterial CtpA, when analyzed in the heterologous chloroplast import assays, showed
transport by both pathways in vitro (Karnauchov et al., 1997). Several studies have shown
that the choice of pathway is dictated by the thylakoid-targeting signal peptides of the
respective transit peptides (Robinson et al., 1994; Henry et al., 1994; Karnauchov et al.,
1994).
Thylakoid transfer signals of the Sec- and ∆pH-dependent proteins share several common
structural features: a hydrophilic, positively charged N-terminal region (n-domain), a
hydrophobic core region (h-domain), and a polar C-terminal region (c-domain). Domain
26
Introduction
swapping experiments and mutagenesis studies showed that pathway specificity is determined
by subtle differences between the signal peptides for the Sec- and ∆pH-dependent pathways
(Henry et al., 1997; Chaddock et al., 1995; Brink et al., 1998). A twin-arginine motif is found
immediately before the h-domain, which is a characteristic of almost all ∆pH-dependent
signal peptides and distinguishes them from those of the Sec-dependent pathway. Sitedirected mutagenesis has shown that this motif plays an important role in pathway
recognition. Even the conservative substitution of a single of these arginines to lysine
dramatically impairs the ability of the precursors to be transported (Chaddock et al., 1995).
Therefore, the ∆pH-dependent pathway was also termed the twin-arginine translocation
pathway (TAT-pathway). Although the RR motif is strictly required for TAT-dependent
translocation, a few natural exceptions have been so far identified, as well as in E.coli (Molik
et al., 2001, Hinsley et al., 2001; Ignatova et al., 2002). At least a KR motif is compelling for
a TAT-dependent transport. Typical signal peptides for the ∆pH/TAT-dependent pathway, as
well as for the Sec-dependent and the spontaneous pathways, are shown in Fig. 2.
∆pH/TAT-dependent pathway
Sp
23K
AQKQDDNEANVLBSGVSRRLALTVLIGAAAVGSKVSPADA
Wh
23K
AQKNDEAASDAAVVTSRRAALSLLAGAAAIAVKVSPAAA
Sp
16K
AQQVSAEAETSRRAMLGFVAAGLASGSFVKAVLA
Ma
16K
ASAEGDAVAQAGRRAVIGLVATGIVGGALSQAARA
Ara
16Ka
AQQSEETSRRSVIGLVAAGLAGGSFVQAVLA
Ara
16Kb
NVSVPESSRRSVIGLVAAGLAGGSFVKAVFA
Bar
PSI-N
Cot
PSII-T
VQMSGERKTEGNNGRREMMFAAAAAAICSVAGVATA
Ara
PSII-T
TPSLEVKEQSSTTMRRDLMFTAAAAAVCSLAKVAMA
Ara
p29
Ara
Hcf136
Ara
p16
Ara
Rieske
...ACQASSIPADRVPDMEKRKTLNLLLLGALSLPTGYMLVPYATFFVPPG
Sp
Rieske
...TCQATSIPADNVPDMQKRETLNLLLLGALSLPTGYMLLPYASFFVPPG
AAAKRVQVAPAKDRRSALLGLAAVFAATAASAGSARA
CSKIEPQVSGESLAFHRRDVLKLAGTAVGMELIGNGFINNVGDAKA
SPSPSSSSSSLSFSRRELLYQSAVSLSLSSIVGPARA
...SKKNQIAYSGNSKNQTSSSLLWKRRELSLGFMSSLVAIGLVSNDRRRHDANA
Sec-dependent pathway
Sp
33K
...SSGGRLSLSLQSDLKELANKCVDATKLAGLALATSALIASGANA
27
Introduction
Wh
33K
AFGVDAGARITCSLQSDIREVASKCADAAKMAGFALATSALLVSGATA
Sp
PC
ASLKNVGAAVVATAAAGLLAGNAMA
Bar
PC
ASLGKKAASAAVAMAAGAMLLGGSAMA
Sp
PSI-F
Bar
PSI-F
Ara
P17.4
...QENDQQQPKKLELAKVGANAAAALALSSVLLSSWSVAPDAAMA
...SGDNNNSTATPSLSASIKTFSAALALSSVLLSSAATSPPPAAA
...SLFPLKELGSIACAALCACTLTIASPVIA
Spontaneous pathway
Sp
PsbW
...PSTTETTTTTNKSMGASLLAAAAAATISNPAMALVDE
Sp
CFoII
...PPLKHLNLSVLKSAAITATPLTLSFLLPYPSLAEEIEK
Sp
PsbS
...KANELFVGRVAMIGFAASLLGEALTGKGILA
Sp
PsbY
ISLSPLGLSNSKLPMGLSPIITAPAIAGAVFATLGSVDPAF
Fig. 2. Signal peptides for ∆ pH/TAT-dependent, Sec-dependent and spontaneous thylakoidal
protein transport pathways. Signal sequences are shown for representative proteins from spinach
(Sp), wheat (Wh), maize (Ma), Arabidopsis (Ara), barley (Ba) and cotton (Cot). The hydrophobic
domains (H-domain) are underlined. The conserved twin-R motif of ∆pH pathway signals (see text)
is shown in bold, as are the positively charged residues found in the n-domains of the Sec-pathway
and spontaneous pathway signal peptides.
Two observations indicate that signal peptides for the ∆pH/TAT-dependent pathway have
specificity determinants in addition to the twin-arginine motif. First, substitution of the RR
motif to KR, RK, or KK is not sufficient to convert a ∆pH-pathway signal peptide to a Sec
type targeting signal peptide (Chaddock et al., 1995). Second, replacing the h/c-domains of a
∆pH-dependent signal peptide by the corresponding domains of a Sec type signal peptide is
tolerable for ∆pH/TAT-dependent pathway recognition, but not vice versa (Henry et al., 1997;
Bogsch et al., 1997), suggesting a Sec-avoidance motif in these regions of ∆pH/TATdependent signal sequences.
ii. Components of the ∆pH/TAT-pathway machinery
Identification of the genes for components of the ∆pH/TAT translocation machinery has
confirmed that this pathway is in fact highly conserved between bacteria and chloroplasts.
28
Introduction
Identification of the components of this pathway was initiated in higher plant by using genetic
approaches. Voelker and Barkan succeeded in isolating a maize mutant hcf106, in which the
translocation of the ∆pH-dependent pathway precursors but not of Sec-pathway precursors
was affected (Voelker and Barkan, 1995). The corresponding gene could be isolated and
sequenced, which led to the identification of the first component of the ∆pH/TAT-dependent
machinery, the Hcf106 protein (Settles et al., 1997). A few years later, a second component of
the ∆pH/TAT-machinery, notably Tha4, was genetically and biochemically identified in
maize (Mori et al., 1999; Walker et al., 1999). Sequencing showed that Hcf106 and Tha4 are
homologous proteins with high similarity in both structure and sequence, especially in the
transmembrane domain (~65% identity). Each protein contains a predicted amino proximal
transmembrane domain, through which these proteins are anchored to thylakoids. Their Cterminal domains vary in sequence and in length, but are both located on the stromal side of
the thylakoid membrane.
As mentioned above, searching for genes homologous gene to hcf106 in the E. coli genome
has led to the identification of the tatABCD operon that encodes the components of a
∆pH/TAT-like system in E. coli. A tatC homology gene has been found in the Arabidopsis
genome (Motohashi et al., 2001), as well as in pea (Mori et al., 2001). CpTatC is located in
the thylakoid membrane, with both N- and C-termini in the stroma. It is predicted to span the
membrane six times. The genes encoding essential ∆pH/TAT pathway components, as well as
the predicted topology of these proteins, are shown in Fig. 3.
29
Introduction
Bacterial Tat genes
yigT
yigT
yigU
yigW
tatA
tatB
tatC
tatD
ybeC
tatE
Plant Tat genes
tha4
hcf106
tatC
N
C
C
C
stroma
Topology
N
Tha4
=TatA
lumen
N
Hcf106
=TatB
cpTatC
=TatC
Fig. 3. ∆ pH/TAT pathway components of bacterial and plant systems. The bacterial TAT
components are encoded by two operons: yig and ybe. Their corresponding ORFs and the original
gene names are indicated. The plant TAT genes are linked to their corresponding bacterial
counterparts by arrows. The schematic topology of each protein in the thylakoid membrane is drawn
according to Mori and Cline (2001).
iii. The mechanism of ∆pH/TAT-dependent protein translocation
The energy requirements of the ∆pH/TAT-dependent pathway are unique among all protein
translocation systems. Transport of precursors by this pathway is independent of nucleoside
triphosphates, which are required in most other systems. Instead, the ∆pH/TAT pathway uses
the transthylakoidal proton gradient to drive transport. Ionophores, such as nigericin or
carbonyl cyanide m-chlorophenylhydrazone (CCCP), inhibit transport. On the other hand,
inhibitors for the Sec-dependent pathway, such as azide and apyrase, have no effect on protein
transported by this pathway (Mould and Robinson, 1991; Cline et al., 1992; Klösgen et al.,
1992). Soluble factors are not important for ∆pH-dependent pathway, suggesting that the
initiative step of the transport takes place at the thylakoid membrane, rather than in the stroma
(Klösgen et al., 1992; Hulford et al., 1994).
30
Introduction
The thylakoidal ∆pH/TAT pathway is able to transport foreign folded proteins that were fused
to a ∆pH/TAT signal peptide (Clark and Theg, 1997, Hynds et al., 1998). However, the
conformation of authentic substrates during transport has not been established, although
indirect evidence suggests that they are folded too. The OE23 (Creighton et al., 1995) and
OE16 (Musser and Theg, 2000) proteins have been shown to fold in the stroma prior to
transport, but it is unclear if they remain folded also during transport. Further evidence comes
from studies on the thylakoidal Rieske protein, a Fe/S containing protein. Τhe Rieske protein
gains its Fe/s cluster presumably in the stroma, which is obligatory for ∆pH/TAT-dependent
translocation (Molik et al., 2001).
The ∆pH/TAT pathway is able to transport proteins ranging in size from 3.6 kDa to at least 80
kDa (Schubert et al., 2001). The largest known substrate in bacteria is formate dehydrogenase
N with a molecular mass of 132 kDa. This protein expands up to 7 nm in diameter when
folded, which is larger than the width of the lipid bilayer (Berks et al., 2000). Although the
system operates in energy-transducing membranes, large-scale translocation by this pathway
appears not to affect the proton permeability (Teter and Theg, 1998). Therefore, the TAT
system must have a quite unique translocation mechanism to prevent the leakage of protons.
A gated but dynamic flexible channel was supposed to be essential for this system (Robinson
et al., 2000). Expansion and contraction of the protein channel could for example be
accomplished by adding or removing component monomers. However, the mechanism by
which a proton gradient can move the precursor protein are not clear as yet.
iv. Capacity of the ∆pH/TAT-pathway
It appears likely that the bacteria and thylakoids have evolved the ∆pH/TAT-pathway to allow
for the translocation of the large folded domains, while preventing ion leakage through the
membranes. A proteome analysis of Arabidopsis thylakoid lumen proteins has been
performed recently (Schubert et al., 2002), suggesting that more than 50% of the lumenal
proteins are synthesized with a typical twin-R motif, indicative of targeting by the ∆pH/TAT
31
Introduction
pathway. In contrast, it has been estimated that in bacteria only about 2.5% of the proteins
exported are substrates of the TAT-pathway, and almost all of them are cofactor-containing
proteins (Berks et al., 2000). Apparently, the number of TAT-substrates has been largely
expanded during the endosymbiotic evolution. Unlike the bacterial TAT proteins, many of the
passenger proteins of the chloroplast ∆pH/TAT-pathway do not carry cofactors. These
proteins include the OE16 and OE23 proteins. For these cofactor-less TAT pathway proteins,
they also appear rapid folding kinetics within the chloroplast stroma. In chloroplast, folding of
passenger proteins prior to thylakoid translocation likely avoid the impediment of the
oscillating acidity of the thylakoid lumen which is directly dependent upon the photosynthetic
activity. Obviously, a folded protein is too large for the Sec system to handle (Robinson et al.,
2000). Thus, the passenger protein of a ∆pH/TAT pathway protein cannot be transported by
the Sec pathway, even when fused to a Sec-type signal peptide (Clausmeyer et al., 1993;
Robinson et al., 1994; Henry et al., 1997; Bogsch et al., 1997). A Sec pathway passenger
protein, in contrast, can be transported by ∆pH pathway when a ∆pH pathway signal is
attached (Clausmeyer et al., 1993; Robinson et al., 1994; Henry et al., 1994).
3.2.4. Spontaneous insertion of thylakoid membrane proteins
Initial studies showed that a range of single-span membrane proteins, including CFoII, PsbW
and PsbX, insert into the thylakoid membrane in the absence of SRP, NTPs, ∆pH or a
functional Sec machinery (Michl et al., 1994; Lorkovic et al., 1995; Kim et al., 1998). Even
pretreatment of the thylakoid membrane with trypsin, which should destroy all stromaexposed domains of the translocases, has no effect on their insertion (Robinson et al., 1996).
A similar mechanism was originally proposed for M13 procoat protein that inserts into the E.
coli plasma membrane by a SRP/Sec-independent pathway (Kuhn et al., 1986). Remarkably,
the precursors of CFoII, PsbW and PsbX largely resemble the M13 procoat precursor in
structure. Like the M13 procoat, they have two hydrophobic domains, one in the signal
peptide, the other in the mature protein. Both the N- and C-terminal domains are positively
charged, whereas the hydrophilic domain between the two hydrophobic domains is negatively
32
Introduction
charged. Like being described for M13 procoat protein (Kuhn, 1987), membrane insertion of
the hydrophobic domains of those proteins leads to translocation of the negatively charged
hydrophilic domain and formation of a loop-like intermediate in the membrane (Thompson et
al., 1998). Therefore, it was previously assumed that a similar spontaneous mechanism is
operating in bacteria and chloroplasts.
C
1.
N+ + +
- - - -
2.
N
3.
C
4.
N
C
N
TPP
Fig. 4. Model for the “spontaneous” insertion of thylakoid membrane proteins. 1. Targeting
of precursor protein to membrane. 2. Formation of hydrophobic α-helixes. 3. Loop insertion and 4.
Cleavage of TPP (modified according to Michl et al., 1994). Positively charges are indicated by
“+”, and hydrophobic α-helixes are drawn as shaded rectangles. TPP is indicated by scissors.
Since insertion of M13 procoat was recently shown to depend on the function of YidC (see
above), it was also assumed that Alb3, the chloroplast homologue to YidC, could be involved
in insertion of the thylakoid membrane proteins. However, pretreatment of the thylakoid
membrane with antibodies against Alb3 strongly inhibits the SRP/Sec-dependent integration
of LHCP, but has no effect on the Sec-independent insertion of CFoII, PsbW and PsbX
(Woolhead et al., 2001). These data, however, have been obtained in experiments in which
the proteins had to be inserted into the isolated thylakoid membrane and therefore may not
necessarily correspond to the situation within the chloroplasts. Thus, it still cannot be ruled
out that a transport apparatus is involved in the “spontaneous” insertion pathway, but so far
there is no positive evidence for this.
33
Introduction
Taken together, four distinct pathways are utilized to transport thylakoid proteins across or
into the thylakoid membrane. Each pathway transports special protein substrates and,
apparently, is subject to differential regulation in biogenesis of the thylakoid membrane.
Possibly, these pathways are needed to avoid catastrophic feedback when demands on protein
translocation are high. A scheme for the multiple export systems of chloroplasts is shown in
Fig. 5.
TOC GTP
cytosol
TIC
outer envelope membrane
ATP
inner envelope membrane
stroma
Spontaneous
SRP564
SRP43
FtsY
SRP
GTP
(∆pH)
no stroma
no NTPs
no ∆pH
TAT
Sec
∆pH
ATP
(∆pH)
TatC
tB
Hcf106
Ta
Tha4
SecY/E
lumen
Alb3
thylakoid
=Stromal Processing Peptidase
=Thylakoidal Processing Peptidase
Fig. 5. Import and sorting of nuclear-encoded thylakoid proteins. Nuclear-encoded thylakoid
proteins are synthesized in precursor form in the cytosol carrying N-terminal transit peptides.
Arrows indicate the path of precursor proteins from the cytosol to the thylakoid membrane. The
precursor proteins are imported into the chloroplast stroma through the TOC-TIC complex driven by
ATP hydrolysis. The stroma targeting sequences are cleaved off by stromal processing peptidase.
The lumenal proteins are translocated by either Sec-dependent or ∆pH/TAT-dependent machineries.
Most of the thylakoid membrane proteins are integrated into the membrane by either SRP-dependent
pathway or spontaneous insertion. Once in the thylakoid membrane or lumen, the signal peptides are
removed by thylakoidal processing peptidase (optionally). Specific requirements of energy and
characteristic factors of each pathway are indicated.
34
Introduction
4. Goal of the work
It was the goal of this work to characterize the mechanism of ∆pH/TAT dependent protein
transport across the thylakoid membrane. For this purpose, in vitro protein transport
experiments were performed using isolated intact chloroplasts (in organello) or thylakoid
vesicles (in thylakoido). In the first part of this work, the phylogenetic relationship of the
bacterial and thylakoidal TAT-transport systems was analyzed in order to obtain insight into
the origin of these systems compare the function of the TAT translocases of both systems. In
these experiments, a bacterial protein notably GFOR (glucose-fructose oxidoreductase) was
analyzed with the heterologous thylakoid system. In the second part of the work, the
translocation steps of ∆pH/TAT-dependent were analyzed by using the chimeric 16/23 protein
as a model protein. This allows for studying the mechanism of each translocation step with
respect to its energy demands and the requirement of proteinaceous membrane components.
In the third part of this work, the involvement of the mature bodies of the chimeric and
authentic substrate proteins in the thylakoid targeting by ∆pH/TAT-pathway was analyzed.
The final part of this work focuses on the structure of the thylakoidal TAT-translocase. Using
a combination of native gel systems and immuno-affinity assays, I have tried to characterize
the role of the three TAT-subunits, TatA, TatB and TatC, with respect to their function and
organization in each step of the ∆pH/TAT-dependent transport process.
35